Straightforward and reliable methods for simultaneously analyzing several constituents of a complex sample are extremely desirable. Polynucleotide arrays (such as DNA or RNA arrays) are known and are used, for example, as diagnostic or screening tools. Such arrays include regions of usually different sequence polynucleotides (“capture agents”) arranged in a predetermined configuration on a support. The arrays are “addressable” in that these regions (sometimes referenced as “array features”) have different predetermined locations (“addresses”) on the support of array. The polynucleotide arrays typically are fabricated on planar supports either by depositing previously obtained polynucleotides onto the support in a site specific fashion or by site specific in situ synthesis of the polynucleotides upon the support. After depositing the polynucleotide capture agents onto the support, the support is typically processed (e.g., washed and blocked for example) and stored prior to use.
In use, an array is contacted with a sample or labeled sample containing analytes (typically, but not necessarily, other polynucleotides) under conditions that promote specific binding of the analytes in the sample to one or more of the capture agents present on the array. Thus, the arrays, when exposed to a sample, will undergo a binding reaction with the sample and exhibit an observed binding pattern. This binding pattern can be detected upon interrogating the array. For example all target polynucleotides (for example, DNA) in the sample can be labeled with a suitable label (such as a fluorescent compound), and the label then can be accurately observed (such as by observing the fluorescence pattern) on the array after exposure of the array to the sample. Assuming that the different sequence polynucleotides were correctly deposited in accordance with the predetermined configuration, then the observed binding pattern will be indicative of the presence and/or concentration of one or more components of the sample. Techniques for scanning arrays are described, for example, in U.S. Pat. No. 5,763,870 and U.S. Pat. No. 5,945,679. Still other techniques useful for observing an array are described in U.S. Pat. No. 5,721,435.
There has been great interest in the analysis of small RNAs, such as short interfering RNAs (siRNAs), microRNAs (miRNA), tiny non-codingRNAs (tncRNA) and small modulatory RNA (smRNA), since the discovery of siRNA biological activity over a decade ago. See Novina et al., Nature 430: 161-164 (2004). Even though the functions of most discovered miRNAs remain a mystery, it has become clear that they exist in abundance in plants and animals, with up to tens of thousands of copies per cell. In the fruit fly, 78 have been identified, and over 200 have been identified in human (see the public database accessible via the website located at >>http://www.sanger.ac.uk/cgi-bin/Rfam/mirna/browse.pl<<). The levels of individual miRNAs seem to vary with developmental stages and tissue types. The level of fluctuation may be correlated with phenotype, mRNA levels, or protein levels for better biological insight. Thus quantitative measurements of miRNA may be of great importance. Further, viral miRNAs have been identified and may play a role in latency (see Pfeffer et al., Science, 304: 734-736 (2004)), making the detection and quantification of miRNAs a potentially valuable diagnostic tool.
Analytic methods employing polynucleotide arrays have been used for investigating these small RNAs, e.g. miRNAs have become a subject of investigation with microarray analysis. See, e.g., Liu et al., Proc. Nat'l Acad. Sci. USA, 101: 9740-9744 (2004); Thomson et al., Nature Methods, 1: 1-7 (2004); and Babak et al., RNA, 10: 1813-1819 (2004). Methods of labeling RNAs are of interest for use in array analysis of RNA to provide an observable label used in interrogating the array. In the study of Liu et al., the miRNA was transcribed into DNA with a biotin-labeled primer. This primer was subsequently labeled with streptavidin-linked Alexa dye prior to array hybridization. This method is susceptible to any reverse-transcriptase reaction bias. Further, the streptavidin-dye as well as streptavidin-biotin-RNA stochiometry may be difficult to quantify. In the study of Thomson et al., the miRNA was directly labeled with 5′-phosphate-cytidyl-uridyl-Cy3-3′ using T4 RNA ligase. This reaction is sensitive to the acceptor sequence. See England et al., Biochemistry, 17: 2069-2776 (1978). In the study of Babak et al (4), the miRNA was labeled with Ulysis Alexa Fluor system, which reacts with guanine residue (G) of RNA. Since different miRNAs do not have uniform G content, this method is not quantitative.
Thus, there is a continuing need for methods of labeling RNA with an observable label. Such methods may be used in conjunction with analytical methods based on observing the label, such as array-based analysis of polynucleotides.